sensors Article
Metal Sulfides as Sensing Materials for Chemoresistive Gas Sensors Andrea Gaiardo 1,2, *, Barbara Fabbri 1,3 , Vincenzo Guidi 1,3,4, *, Pierluigi Bellutti 2 , Alessio Giberti 4 , Sandro Gherardi 1 , Lia Vanzetti 2 , Cesare Malagù 1,3 and Giulia Zonta 1,3 1 2 3 4
*
Department of Physics and Earth Science, University of Ferrara, Via Saragat 1/c, Ferrara 44122, Italy; [email protected] (B.F.); [email protected] (S.G.); [email protected] (C.M.); [email protected] (G.Z.) MNF- Micro Nano Facility, Bruno Kessler Foundation, Via Sommarive 18, Trento 38123, Italy; [email protected] (P.B.); [email protected] (L.V.) CNR-INO—Istituto Nazionale di Ottica, Largo Enrico Fermi 6, Firenze 50124, Italy MIST E-R s.c.r.l., Via P. Gobetti 101, Bologna 40129, Italy; [email protected] Correspondence: [email protected] (A.G.); [email protected] (V.G.); Tel.: +39-0532-974-210 (V.G.); Fax: +39-0532-974-205 (V.G.)
Academic Editor: Michael Tiemann Received: 24 September 2015; Accepted: 22 February 2016; Published: 26 February 2016
Abstract: This work aims at a broad overview of the results obtained with metal-sulfide materials in the field of chemoresistive gas sensing. Indeed, despite the well-known electrical, optical, structural and morphological features previously described in the literature, metal sulfides present lack of investigation for gas sensing applications, a field in which the metal oxides still maintain a leading role owing to their high sensitivity, low cost, small dimensions and simple integration, in spite of the wide assortment of sensing materials. However, despite their great advantages, metal oxides have shown significant drawbacks, which have led to the search for new materials for gas sensing devices. In this work, Cadmium Sulfide and Tin (IV) Sulfide were investigated as functional materials for thick-film chemoresistive gas-sensors fabrication and they were tested both in thermo- and in photo-activation modes. Furthermore, electrical characterization was carried out in order to verify their gas sensing properties and material stability, by comparing the results obtained with metal sulfides to those obtained by using their metal-oxides counterparts. The results highlighted the possibility to use metal sulfides as a novel class of sensing materials, owing to their selectivity to specific compounds, stability, and the possibility to operate at room temperature. Keywords: metals sulfides; chemoresistive gas sensors; thick-film; cadmium sulfide; tin (IV) sulfide
1. Introduction The great challenge of nanostructured materials lies in the control of their properties by the grain size, which combines bulk and surface effects [1–4]. Low-dimensional nanostructures have been prepared with various morphologies and have attracted research attention because of their fundamental role in the comprehension of the quantum size effect and great potential applications [5,6]. One-dimensional (1D) nanostructures are ideal for investigating the dependence of electrical transport, mechanical and optical properties on size and dimensionality [7]. Indeed, highly attractive properties and novel applications have resulted from well-aligned one-dimensional nanostructures on substrates, because they play a key role as both interconnections and functional components in improving performance of technologically advanced devices [8,9]. In recent years, many unique and excellent properties have already been demonstrated or proposed, such as superior mechanical toughness, lower turn-on voltage for field emitters, higher efficiency for solar cells, better electrochemical performance Sensors 2016, 16, 296; doi:10.3390/s16030296
www.mdpi.com/journal/sensors
Sensors 2016, 16, 296
2 of 19
for lithium-ion batteries and enhancement of thermoelectric figure of merit [10,11]. At the same time, two-dimensional (2D) nanostructures, i.e., nanosheets, nanoplates, and nanowalls, are suggested to be ideal components for nanoscale devices used in data storage, nanoswitches and biological sensors, due to their nanometre-scale thickness, high surface-to-volume ratio, and fascinating photocatalytic and optical activities [12]. In the last years, the variable features of colloidal nanocrystals, such as their size-dependent electronic, optical, magnetic, mechanical and chemical properties, which cannot be obtained in their bulk counterparts, have attracted the attention of researchers [13,14]. Within colloidal semiconductors, metal chalcogenide nanocrystals have been extensively investigated due to their size-dependent photoemission characteristics and quantum confinement effects [15]. These nanomaterials can be used for different biological labelling and diagnostics, electroluminescent devices, lasers, photovoltaic devices, light-emitting diodes and single-electron transistors [16]. Among colloidal nanocrystals, metal oxides have gained a significant role in technology development due to their exceptional skills. In recent times, several research works have been focused on these semiconductors to explore new application fields, such as optical, electronic, optoelectronic and biological domains. In particular, the application in which metal oxides have been widely used is gas sensing. The performance of sensors based on metal oxides depends crucially on their dimensions, morphology, composition and surface activity [17]. Among the several parameters that influence the sensing properties of a metal oxide, the potential barrier at the interface between grains is a major physical quantity [18]. Indeed, in this sense, the broad assortment of one-, two- and three-dimensional metal-oxides nanostructures has been a precious source for gas sensors technology, which owes its constant development to the requirements of physical, chemical and biological detection systems [19–22]. Metal sulfides are nanocrystals with great potential for investigation, due to their various types of structures. They are abundant and cheap because they exist in nature as minerals, i.e., heazlewoodite (Ni3 S2 ), chalcocite (Cu2 S), pyrite (FeS2 ) and others. The morphology of metal-sulfide nanostructures can be controlled by applying general solution methods and thermal evaporations, and their possible applications in energy conversion and storage were demonstrated. In the scientific literature, many papers have been reported to provide an overview of recent research and significant advances, ranging from synthesis to properties and applications, especially in energy conversion and storage, such as solar cells, lithium-ion batteries, piezoelectric nanogenerators and fuel-cells [23–25]. So far, in the gas sensing field, metal sulfides have been mainly studied in combination with metal oxides in order to modify the sensing activity of the latter [26,27]. Metal sulfides as sensing materials for gas detection have been poorly studied, and the works published do not present an in-depth study about their sensing properties [28,29]. On the contrary, the literature presents extensive investigations on metal-oxide semiconductors as sensing materials, due to their excellent sensitivity, fast response and recovery times, and low-cost [30,31]. However, despite such important advantages, metal-oxides still exhibits unsolved drawbacks. Their incomplete selectivity and lack of stability sometimes result in unreliable responses [32,33]. Moreover, these semiconductors often need a significant amount of energy to support chemical reactions at the surface, activated at high temperatures. By studying physical and chemical properties of nanostructured metal sulfides, it arose that such materials may be very good candidates to be further investigated in the chemoresistive gas sensing field. Indeed, by using these materials, we expect an improvement from an energy consumption point of view, both in thermaland photo-activation modes, due to their lower band-gap than for metal-oxide semiconductors. This means that the activation of intrinsic surface reactions occurs at lower working temperatures, then minor power supply is necessary. Due to this advantage, we were motivated in the search for potential improved performance in terms of selectivity and stability. The absence of oxygen in the crystal lattice of metal sulfides leads to a different catalytic mechanism on the surface reaction with respect to metal oxides. In addition, this absence may solve the constant drift of the signal suffered by metal oxides and ascribed to the in/out diffusion of oxygen vacancies, which alters the doping level. For these reasons,
Sensors 2016, 16, 296
3 of 19
we decided to focus our work on the use of metal sulfides for chemoresistive gas sensors by means of thick-film deposition technique. In this work, the sensing properties of Cadmium Sulfide (CdS) and Tin (IV) Sulfide (SnS2 ) were deeply studied in thermo- and photo-activation mode. Some characteristics of these two metal sulfides, such as the possibility to synthesize them on a nanometric scale with simple and inexpensive methods and their thermal stability, encouraged their usage as sensing films. Moreover, each of these metal sulfides is the counterpart of a metal oxide, in particular, SnS2 is analogous to the widely used SnO2 . Hence, a comparison of the sensing properties and performance between metal sulfides and their metal-oxide counterparts was carried out, in order to identify the main promising features of these semiconductors as gas sensing materials. 2. Experimental Section 2.1. Chemical Synthesis and Thick-Films Deposition Cadmium Sulfide (CdS) and Tin (IV) Sulfide (SnS2 ) were synthesized as nanoparticles via precipitation reactions in aqueous solution, using thioacetamide as a source of S2´ ions and metalorganic or salt compounds to release metal ions. The chemical reagents used in these syntheses were from Sigma Aldrich. In order to control the growth of crystals, which was expected to be nanometric, a complex agent was used and chemicals adapted to adjust the solution pH. 2.1.1. Synthesis of Cadmium Sulfide Cadmium Sulfide nanoparticles were obtained by precipitation method at room temperature and atmospheric pressure, in aqueous solution. In this synthesis, 10 mmol of cadmium acetate dihydrate and 20 mmol of o-phenylenediamine were dissolved in 100 mL of water and stirred for 2 h. Afterwards, 20 mmol of thioacetamide were added to the solution. The mixture obtained was stirred for 6 h. Hence, a precipitate of yellow-orange nanoparticles was formed. The product was isolated by vacuum filtration and washed several times with methanol and water. At last, CdS nanoparticles were dried for 4 h at 40 ˝ C. The synthesis was performed in three different modes: without o-phenylenediamine (sample S1), with 10 mmol (sample S2), and with 20 mmol of o-phenylenediamine as complexing agent (sample S3). 2.1.2. Synthesis of Tin (IV) Sulfide The Tin (IV) Sulfide nanoparticles were synthesized through precipitation at standard pressure and temperature, in aqueous solution. First, 1.65 mmol of SnCl4 ¨ 5H2 O were dissolved in a beaker with HCl (37% m/v). Then, to the resulting suspension were added distilled water, diluted in 80 mL. The solution was stirred for 10 min. Afterwards, 0.25 g of thioacetamide and 20 mL were added to this solution. The mixture obtained was stirred for further 3 h. The Tin (IV) Sulfide precipitated in this solution as brown nanoparticles, thus it was isolated by vacuum filtration and washed with water and methanol. At last, the product was dried for 6 h at 40 ˝ C. Also for SnS2, three synthesis were carried out: without HCl (sample ST1), with HCl as chemical to adjust pH (pH = 3) (sample ST2), and with the same quantity of HCl (pH = 3) and 3.3 mmol of o-phenylenediamine as complexing agent (sample ST3). Organic vehicles were added to CdS and SnS2 nanopowders in order to obtain pastes with a suitable viscosity, to allow the deposition of the sensing layers onto alumina substrates through the screen printing technique (thickness„30 µm) [34]. In particular, the organic vehicle used was composed of a glycol ether as wetting agent and an acrylic resin. The front-side of alumina substrates provides interdigitated Au electrodes for the measure of the film resistance, while the back-side is equipped with a heater to apply the optimal working temperature of the sensors. Afterwards, to obtain the thermal stabilization, the screen-printed films were treated at 180 ˝ C in a muffle oven for 12 h in air,
Sensors 2016, 16, 296
4 of 19
allowing the evaporation of organic vehicles. The substrates were finally bonded on a suitable support Sensors 2016, 16, 296 4 of 18 to be connected with the electronic system (Figure 1a).
Figure Figure 1. 1. (a) (a) Image Image of of the the gas gas sensor sensor device device and and (b) (b) the the schematic schematic representation representation of of the the gas gas sensing sensing system used for electrical characterization in photo-activation mode. system used for electrical characterization in photo-activation mode.
2.2. 2.2. Chemical, Chemical, Morphological Morphological and and Structural Structural Characterizations Characterizations Powders and films were were studied studied with with Energy Energy Dispersive Dispersive X-Ray X-Ray spectroscopy spectroscopy and and Scanning Scanning Electron Microscopy (SEM-EDX spectroscopy) techniques, to investigate morphology and chemical (SEM-EDX investigate morphology chemical composition composition of of the obtained obtained materials. materials. The instrument used was a Zeiss EVO 40 Microscope with an acceleration voltage voltage of of 30 30 kV. kV. Further information on powders were by TEM images, obtained by a Hitachi H800 microscope, supplied with with aa tungsten tungsten gun gun with with maximum maximum voltage voltage of of 200 200 kV. kV. X-Ray Diffraction (XRD) analysis was carried out on analysis carried out on the the as-synthesized as-synthesized CdS CdS and and SnS SnS22 nanopowders and on the the thermal thermal processed processed powders. powders. The instrument was a Bruker D8 D8 Advance Advance diffractometer equipped with a Si(Li) solid-state solid-state detector(SOL-X) set to measure CuKα 1,2 radiation and with an X-ray tube operating at 40 kVkV andand 40 mA. An alumina and and zero zero background holders were and with an X-ray tube operating at 40 40 mA. An alumina background holders ˝ used sideasloaded for thefor samples, respectively. Measuring conditions were 5–95 2θ range, 0.02˝ were as used side loaded the samples, respectively. Measuring conditions were 5–95° 2θ range, 2θ scan counting time pertime step 4per andstep 6 s for as synthesized thermallyand treated nanopowders, 0.02° 2θrate, scan rate, counting 4 and 6 s for as and synthesized thermally treated respectively. The phase identification was achieved by was search-match the EVA v.14.0 program by nanopowders, respectively. The phase identification achieved using by search-match using the EVA Bruker and the Powder Diffraction File database (PDF) File v. 9.0.133. To define crystallite of the v.14.0 program by Bruker and the Powder Diffraction database (PDF) the v. 9.0.133. To size define nanopowders, v.4.1 program used,v.4.1 basedprogram on the Rietveld method and crystallite size the of TOPAS the nanopowders, thewas TOPAS was used, based onaccomplished the Rietveld by the Double–Voigt approachby[35,36]. The line-profile fitting was obtained the fundamental method and accomplished the Double–Voigt approach [35,36]. Thethrough line-profile fitting was parameters [37–39]. obtained through the fundamental parameters [37–39]. For XPS measurements, measurements, the powders powders were attached attached to the the sample sample holder holder using using aa double-sided double-sided carbon tape. XPS spectra spectra were were recorded recorded using using aa Scienta Scienta Esca-200 Esca-200 system equipped with a tape. XPS monochromatized Al Ka (1486.6 eV) source. An overall overall energy resolution of 0.4 eV was routinely used. The emission angle between the axis of the analyzer and the normal to the sample surface was negligible. 1s core core levels levels were were collected. collected. Charge Charge compensation negligible. For For each each sample sample Sn Sn 3d, S 2p, O 1s and C 1s was achieved using a flood gun and all core level peak energies were referenced was achieved flood gun and all core level peak energies were referenced to to the the saturated saturated hydrocarbon hydrocarbon in in C C 11 ss at at 285.0 285.0 eV. eV.
The thermogravimetric analysis (TG/DTG/DTA) of the sensing materials were carried out using a Netzsch 409 PC Luxx TG/DTA thermal analyzer. A proper amount of samples were filled in a nickel crucible and analyzed in the range 20–800 °C, with a heating rate of 10 °C·min−1 under air flow of 20 mL·h−1. UV–vis absorption measurements were performed to investigate the optoelectronic properties of the synthesized semiconductor, by using a Cary 50 Varion instrument in the range 300–900 nm
Sensors 2016, 16, 296
5 of 19
The thermogravimetric analysis (TG/DTG/DTA) of the sensing materials were carried out using a Netzsch 409 PC Luxx TG/DTA thermal analyzer. A proper amount of samples were filled in a nickel crucible and analyzed in the range 20–800 ˝ C, with a heating rate of 10 ˝ C¨min´1 under air flow of 20 mL¨h´1 . UV–vis absorption measurements were performed to investigate the optoelectronic properties of the synthesized semiconductor, by using a Cary 50 Varion instrument in the range 300–900 nm (Virtual double-radius). The dimethyl sulfoxide was used as reference and solvent. The cuvette was made of quartz and the optical path was 1 cm. 2.3. Gas Sensing Measurements SnS2 and CdS sensors were electrically characterized in a dedicated chamber for gas measurements by means of the flow-through technique. The sensors were heated at their working temperature under a continuous flow of synthetic air for a few hours before testing the gases, in order to achieve the thermodynamic equilibrium of the SnS2 and CdS grains on the surface. Air and gases were from certified bottles and their injection in the chamber was carried out by means of a PC-driven mass-flow-controller. The conductance of the films was constantly recorded during the gas measurements through proper electronics interfaced to a data-acquiring system. For a n-type semiconductor, the responses to reducing and oxidizing agents were calculated as: # ` ˘ Ggas ´ Gair {Gair f or reducing gases ` ˘ “ Ggas ´ Gair {Ggas f or oxidizing gases
(1)
where Ggas and Gair are the conductance values in gas and in air, respectively. 2.3.1. Gas Measurements in Thermo-Activation Mode The performance of the sensing films was investigated at operating temperatures ranging between 150 ˝ C and 300 ˝ C for SnS2 -based sensors, and between 150 ˝ C and 350 ˝ C for CdS-based sensors. Higher temperatures must be avoided with these materials because they oxidize at temperatures of 400 ˝ C and 500 ˝ C for SnS2 and CdS, respectively [40–42]. For this reason, we decided to test these sensors at temperature fairly lower than their critical temperature. Both the sensors highlighted, at temperatures lower than 250 ˝ C, an extremely low chemoresistive activity. This result proved that temperatures lower than 250 ˝ C are not sufficient to activate chemical reactions at the surface. The tested gases represent different categories of molecules, and, in this way, it was possible to verify the surface reactivity of these semiconductors with respect to analytes characterized by important chemical differences. Gas concentrations were chosen taking into account the corresponding Threshold Limit Value (TLV) [43]. 2.3.2. Arrhenius Plot and Intergrain Barrier Measurements Arrhenius plot and intergrain barrier vs. temperature measurements were performed to compare the behavior of metal sulfides with their corresponding metal oxides. The analysis was conducted on the thick films at temperatures ranging within 100–500 ˝ C. All measurements were carried out in a sealed chamber at 25 ˝ C under atmosphere controlled with a constant flow rate (0.5 L/min) of synthetic air [44]. 2.3.3. Gas Measurements in Photo-Activation Mode For the electrical characterization in photo-activation mode, the sensors were placed in a dedicated chamber provided with a glass window, through which the light emitted by a Light Emitting Diode (LED) was focused onto the films by means of an optical system (Figure 1b). The LEDs were quasi-monochromatic with an emission spectrum width of 5 nm. Synthetic air and gases were injected into the chamber by the flow-through technique. The electrical conductance of the films was recorded
Sensors 2016, 16, 296
6 of 19
continuously during the experiments by using a simple circuit based on an operational amplifier. The films were maintained under a continuous flow of synthetic dry air during all experiments. At first, CdS and SnS2 sensors were tested in dark condition. In this case, we observed that the conductance was too low to be measured with our experimental setup (
Metal Sulfides as Sensing Materials for Chemoresistive Gas Sensors Andrea Gaiardo 1,2, *, Barbara Fabbri 1,3 , Vincenzo Guidi 1,3,4, *, Pierluigi Bellutti 2 , Alessio Giberti 4 , Sandro Gherardi 1 , Lia Vanzetti 2 , Cesare Malagù 1,3 and Giulia Zonta 1,3 1 2 3 4
*
Department of Physics and Earth Science, University of Ferrara, Via Saragat 1/c, Ferrara 44122, Italy; [email protected] (B.F.); [email protected] (S.G.); [email protected] (C.M.); [email protected] (G.Z.) MNF- Micro Nano Facility, Bruno Kessler Foundation, Via Sommarive 18, Trento 38123, Italy; [email protected] (P.B.); [email protected] (L.V.) CNR-INO—Istituto Nazionale di Ottica, Largo Enrico Fermi 6, Firenze 50124, Italy MIST E-R s.c.r.l., Via P. Gobetti 101, Bologna 40129, Italy; [email protected] Correspondence: [email protected] (A.G.); [email protected] (V.G.); Tel.: +39-0532-974-210 (V.G.); Fax: +39-0532-974-205 (V.G.)
Academic Editor: Michael Tiemann Received: 24 September 2015; Accepted: 22 February 2016; Published: 26 February 2016
Abstract: This work aims at a broad overview of the results obtained with metal-sulfide materials in the field of chemoresistive gas sensing. Indeed, despite the well-known electrical, optical, structural and morphological features previously described in the literature, metal sulfides present lack of investigation for gas sensing applications, a field in which the metal oxides still maintain a leading role owing to their high sensitivity, low cost, small dimensions and simple integration, in spite of the wide assortment of sensing materials. However, despite their great advantages, metal oxides have shown significant drawbacks, which have led to the search for new materials for gas sensing devices. In this work, Cadmium Sulfide and Tin (IV) Sulfide were investigated as functional materials for thick-film chemoresistive gas-sensors fabrication and they were tested both in thermo- and in photo-activation modes. Furthermore, electrical characterization was carried out in order to verify their gas sensing properties and material stability, by comparing the results obtained with metal sulfides to those obtained by using their metal-oxides counterparts. The results highlighted the possibility to use metal sulfides as a novel class of sensing materials, owing to their selectivity to specific compounds, stability, and the possibility to operate at room temperature. Keywords: metals sulfides; chemoresistive gas sensors; thick-film; cadmium sulfide; tin (IV) sulfide
1. Introduction The great challenge of nanostructured materials lies in the control of their properties by the grain size, which combines bulk and surface effects [1–4]. Low-dimensional nanostructures have been prepared with various morphologies and have attracted research attention because of their fundamental role in the comprehension of the quantum size effect and great potential applications [5,6]. One-dimensional (1D) nanostructures are ideal for investigating the dependence of electrical transport, mechanical and optical properties on size and dimensionality [7]. Indeed, highly attractive properties and novel applications have resulted from well-aligned one-dimensional nanostructures on substrates, because they play a key role as both interconnections and functional components in improving performance of technologically advanced devices [8,9]. In recent years, many unique and excellent properties have already been demonstrated or proposed, such as superior mechanical toughness, lower turn-on voltage for field emitters, higher efficiency for solar cells, better electrochemical performance Sensors 2016, 16, 296; doi:10.3390/s16030296
www.mdpi.com/journal/sensors
Sensors 2016, 16, 296
2 of 19
for lithium-ion batteries and enhancement of thermoelectric figure of merit [10,11]. At the same time, two-dimensional (2D) nanostructures, i.e., nanosheets, nanoplates, and nanowalls, are suggested to be ideal components for nanoscale devices used in data storage, nanoswitches and biological sensors, due to their nanometre-scale thickness, high surface-to-volume ratio, and fascinating photocatalytic and optical activities [12]. In the last years, the variable features of colloidal nanocrystals, such as their size-dependent electronic, optical, magnetic, mechanical and chemical properties, which cannot be obtained in their bulk counterparts, have attracted the attention of researchers [13,14]. Within colloidal semiconductors, metal chalcogenide nanocrystals have been extensively investigated due to their size-dependent photoemission characteristics and quantum confinement effects [15]. These nanomaterials can be used for different biological labelling and diagnostics, electroluminescent devices, lasers, photovoltaic devices, light-emitting diodes and single-electron transistors [16]. Among colloidal nanocrystals, metal oxides have gained a significant role in technology development due to their exceptional skills. In recent times, several research works have been focused on these semiconductors to explore new application fields, such as optical, electronic, optoelectronic and biological domains. In particular, the application in which metal oxides have been widely used is gas sensing. The performance of sensors based on metal oxides depends crucially on their dimensions, morphology, composition and surface activity [17]. Among the several parameters that influence the sensing properties of a metal oxide, the potential barrier at the interface between grains is a major physical quantity [18]. Indeed, in this sense, the broad assortment of one-, two- and three-dimensional metal-oxides nanostructures has been a precious source for gas sensors technology, which owes its constant development to the requirements of physical, chemical and biological detection systems [19–22]. Metal sulfides are nanocrystals with great potential for investigation, due to their various types of structures. They are abundant and cheap because they exist in nature as minerals, i.e., heazlewoodite (Ni3 S2 ), chalcocite (Cu2 S), pyrite (FeS2 ) and others. The morphology of metal-sulfide nanostructures can be controlled by applying general solution methods and thermal evaporations, and their possible applications in energy conversion and storage were demonstrated. In the scientific literature, many papers have been reported to provide an overview of recent research and significant advances, ranging from synthesis to properties and applications, especially in energy conversion and storage, such as solar cells, lithium-ion batteries, piezoelectric nanogenerators and fuel-cells [23–25]. So far, in the gas sensing field, metal sulfides have been mainly studied in combination with metal oxides in order to modify the sensing activity of the latter [26,27]. Metal sulfides as sensing materials for gas detection have been poorly studied, and the works published do not present an in-depth study about their sensing properties [28,29]. On the contrary, the literature presents extensive investigations on metal-oxide semiconductors as sensing materials, due to their excellent sensitivity, fast response and recovery times, and low-cost [30,31]. However, despite such important advantages, metal-oxides still exhibits unsolved drawbacks. Their incomplete selectivity and lack of stability sometimes result in unreliable responses [32,33]. Moreover, these semiconductors often need a significant amount of energy to support chemical reactions at the surface, activated at high temperatures. By studying physical and chemical properties of nanostructured metal sulfides, it arose that such materials may be very good candidates to be further investigated in the chemoresistive gas sensing field. Indeed, by using these materials, we expect an improvement from an energy consumption point of view, both in thermaland photo-activation modes, due to their lower band-gap than for metal-oxide semiconductors. This means that the activation of intrinsic surface reactions occurs at lower working temperatures, then minor power supply is necessary. Due to this advantage, we were motivated in the search for potential improved performance in terms of selectivity and stability. The absence of oxygen in the crystal lattice of metal sulfides leads to a different catalytic mechanism on the surface reaction with respect to metal oxides. In addition, this absence may solve the constant drift of the signal suffered by metal oxides and ascribed to the in/out diffusion of oxygen vacancies, which alters the doping level. For these reasons,
Sensors 2016, 16, 296
3 of 19
we decided to focus our work on the use of metal sulfides for chemoresistive gas sensors by means of thick-film deposition technique. In this work, the sensing properties of Cadmium Sulfide (CdS) and Tin (IV) Sulfide (SnS2 ) were deeply studied in thermo- and photo-activation mode. Some characteristics of these two metal sulfides, such as the possibility to synthesize them on a nanometric scale with simple and inexpensive methods and their thermal stability, encouraged their usage as sensing films. Moreover, each of these metal sulfides is the counterpart of a metal oxide, in particular, SnS2 is analogous to the widely used SnO2 . Hence, a comparison of the sensing properties and performance between metal sulfides and their metal-oxide counterparts was carried out, in order to identify the main promising features of these semiconductors as gas sensing materials. 2. Experimental Section 2.1. Chemical Synthesis and Thick-Films Deposition Cadmium Sulfide (CdS) and Tin (IV) Sulfide (SnS2 ) were synthesized as nanoparticles via precipitation reactions in aqueous solution, using thioacetamide as a source of S2´ ions and metalorganic or salt compounds to release metal ions. The chemical reagents used in these syntheses were from Sigma Aldrich. In order to control the growth of crystals, which was expected to be nanometric, a complex agent was used and chemicals adapted to adjust the solution pH. 2.1.1. Synthesis of Cadmium Sulfide Cadmium Sulfide nanoparticles were obtained by precipitation method at room temperature and atmospheric pressure, in aqueous solution. In this synthesis, 10 mmol of cadmium acetate dihydrate and 20 mmol of o-phenylenediamine were dissolved in 100 mL of water and stirred for 2 h. Afterwards, 20 mmol of thioacetamide were added to the solution. The mixture obtained was stirred for 6 h. Hence, a precipitate of yellow-orange nanoparticles was formed. The product was isolated by vacuum filtration and washed several times with methanol and water. At last, CdS nanoparticles were dried for 4 h at 40 ˝ C. The synthesis was performed in three different modes: without o-phenylenediamine (sample S1), with 10 mmol (sample S2), and with 20 mmol of o-phenylenediamine as complexing agent (sample S3). 2.1.2. Synthesis of Tin (IV) Sulfide The Tin (IV) Sulfide nanoparticles were synthesized through precipitation at standard pressure and temperature, in aqueous solution. First, 1.65 mmol of SnCl4 ¨ 5H2 O were dissolved in a beaker with HCl (37% m/v). Then, to the resulting suspension were added distilled water, diluted in 80 mL. The solution was stirred for 10 min. Afterwards, 0.25 g of thioacetamide and 20 mL were added to this solution. The mixture obtained was stirred for further 3 h. The Tin (IV) Sulfide precipitated in this solution as brown nanoparticles, thus it was isolated by vacuum filtration and washed with water and methanol. At last, the product was dried for 6 h at 40 ˝ C. Also for SnS2, three synthesis were carried out: without HCl (sample ST1), with HCl as chemical to adjust pH (pH = 3) (sample ST2), and with the same quantity of HCl (pH = 3) and 3.3 mmol of o-phenylenediamine as complexing agent (sample ST3). Organic vehicles were added to CdS and SnS2 nanopowders in order to obtain pastes with a suitable viscosity, to allow the deposition of the sensing layers onto alumina substrates through the screen printing technique (thickness„30 µm) [34]. In particular, the organic vehicle used was composed of a glycol ether as wetting agent and an acrylic resin. The front-side of alumina substrates provides interdigitated Au electrodes for the measure of the film resistance, while the back-side is equipped with a heater to apply the optimal working temperature of the sensors. Afterwards, to obtain the thermal stabilization, the screen-printed films were treated at 180 ˝ C in a muffle oven for 12 h in air,
Sensors 2016, 16, 296
4 of 19
allowing the evaporation of organic vehicles. The substrates were finally bonded on a suitable support Sensors 2016, 16, 296 4 of 18 to be connected with the electronic system (Figure 1a).
Figure Figure 1. 1. (a) (a) Image Image of of the the gas gas sensor sensor device device and and (b) (b) the the schematic schematic representation representation of of the the gas gas sensing sensing system used for electrical characterization in photo-activation mode. system used for electrical characterization in photo-activation mode.
2.2. 2.2. Chemical, Chemical, Morphological Morphological and and Structural Structural Characterizations Characterizations Powders and films were were studied studied with with Energy Energy Dispersive Dispersive X-Ray X-Ray spectroscopy spectroscopy and and Scanning Scanning Electron Microscopy (SEM-EDX spectroscopy) techniques, to investigate morphology and chemical (SEM-EDX investigate morphology chemical composition composition of of the obtained obtained materials. materials. The instrument used was a Zeiss EVO 40 Microscope with an acceleration voltage voltage of of 30 30 kV. kV. Further information on powders were by TEM images, obtained by a Hitachi H800 microscope, supplied with with aa tungsten tungsten gun gun with with maximum maximum voltage voltage of of 200 200 kV. kV. X-Ray Diffraction (XRD) analysis was carried out on analysis carried out on the the as-synthesized as-synthesized CdS CdS and and SnS SnS22 nanopowders and on the the thermal thermal processed processed powders. powders. The instrument was a Bruker D8 D8 Advance Advance diffractometer equipped with a Si(Li) solid-state solid-state detector(SOL-X) set to measure CuKα 1,2 radiation and with an X-ray tube operating at 40 kVkV andand 40 mA. An alumina and and zero zero background holders were and with an X-ray tube operating at 40 40 mA. An alumina background holders ˝ used sideasloaded for thefor samples, respectively. Measuring conditions were 5–95 2θ range, 0.02˝ were as used side loaded the samples, respectively. Measuring conditions were 5–95° 2θ range, 2θ scan counting time pertime step 4per andstep 6 s for as synthesized thermallyand treated nanopowders, 0.02° 2θrate, scan rate, counting 4 and 6 s for as and synthesized thermally treated respectively. The phase identification was achieved by was search-match the EVA v.14.0 program by nanopowders, respectively. The phase identification achieved using by search-match using the EVA Bruker and the Powder Diffraction File database (PDF) File v. 9.0.133. To define crystallite of the v.14.0 program by Bruker and the Powder Diffraction database (PDF) the v. 9.0.133. To size define nanopowders, v.4.1 program used,v.4.1 basedprogram on the Rietveld method and crystallite size the of TOPAS the nanopowders, thewas TOPAS was used, based onaccomplished the Rietveld by the Double–Voigt approachby[35,36]. The line-profile fitting was obtained the fundamental method and accomplished the Double–Voigt approach [35,36]. Thethrough line-profile fitting was parameters [37–39]. obtained through the fundamental parameters [37–39]. For XPS measurements, measurements, the powders powders were attached attached to the the sample sample holder holder using using aa double-sided double-sided carbon tape. XPS spectra spectra were were recorded recorded using using aa Scienta Scienta Esca-200 Esca-200 system equipped with a tape. XPS monochromatized Al Ka (1486.6 eV) source. An overall overall energy resolution of 0.4 eV was routinely used. The emission angle between the axis of the analyzer and the normal to the sample surface was negligible. 1s core core levels levels were were collected. collected. Charge Charge compensation negligible. For For each each sample sample Sn Sn 3d, S 2p, O 1s and C 1s was achieved using a flood gun and all core level peak energies were referenced was achieved flood gun and all core level peak energies were referenced to to the the saturated saturated hydrocarbon hydrocarbon in in C C 11 ss at at 285.0 285.0 eV. eV.
The thermogravimetric analysis (TG/DTG/DTA) of the sensing materials were carried out using a Netzsch 409 PC Luxx TG/DTA thermal analyzer. A proper amount of samples were filled in a nickel crucible and analyzed in the range 20–800 °C, with a heating rate of 10 °C·min−1 under air flow of 20 mL·h−1. UV–vis absorption measurements were performed to investigate the optoelectronic properties of the synthesized semiconductor, by using a Cary 50 Varion instrument in the range 300–900 nm
Sensors 2016, 16, 296
5 of 19
The thermogravimetric analysis (TG/DTG/DTA) of the sensing materials were carried out using a Netzsch 409 PC Luxx TG/DTA thermal analyzer. A proper amount of samples were filled in a nickel crucible and analyzed in the range 20–800 ˝ C, with a heating rate of 10 ˝ C¨min´1 under air flow of 20 mL¨h´1 . UV–vis absorption measurements were performed to investigate the optoelectronic properties of the synthesized semiconductor, by using a Cary 50 Varion instrument in the range 300–900 nm (Virtual double-radius). The dimethyl sulfoxide was used as reference and solvent. The cuvette was made of quartz and the optical path was 1 cm. 2.3. Gas Sensing Measurements SnS2 and CdS sensors were electrically characterized in a dedicated chamber for gas measurements by means of the flow-through technique. The sensors were heated at their working temperature under a continuous flow of synthetic air for a few hours before testing the gases, in order to achieve the thermodynamic equilibrium of the SnS2 and CdS grains on the surface. Air and gases were from certified bottles and their injection in the chamber was carried out by means of a PC-driven mass-flow-controller. The conductance of the films was constantly recorded during the gas measurements through proper electronics interfaced to a data-acquiring system. For a n-type semiconductor, the responses to reducing and oxidizing agents were calculated as: # ` ˘ Ggas ´ Gair {Gair f or reducing gases ` ˘ “ Ggas ´ Gair {Ggas f or oxidizing gases
(1)
where Ggas and Gair are the conductance values in gas and in air, respectively. 2.3.1. Gas Measurements in Thermo-Activation Mode The performance of the sensing films was investigated at operating temperatures ranging between 150 ˝ C and 300 ˝ C for SnS2 -based sensors, and between 150 ˝ C and 350 ˝ C for CdS-based sensors. Higher temperatures must be avoided with these materials because they oxidize at temperatures of 400 ˝ C and 500 ˝ C for SnS2 and CdS, respectively [40–42]. For this reason, we decided to test these sensors at temperature fairly lower than their critical temperature. Both the sensors highlighted, at temperatures lower than 250 ˝ C, an extremely low chemoresistive activity. This result proved that temperatures lower than 250 ˝ C are not sufficient to activate chemical reactions at the surface. The tested gases represent different categories of molecules, and, in this way, it was possible to verify the surface reactivity of these semiconductors with respect to analytes characterized by important chemical differences. Gas concentrations were chosen taking into account the corresponding Threshold Limit Value (TLV) [43]. 2.3.2. Arrhenius Plot and Intergrain Barrier Measurements Arrhenius plot and intergrain barrier vs. temperature measurements were performed to compare the behavior of metal sulfides with their corresponding metal oxides. The analysis was conducted on the thick films at temperatures ranging within 100–500 ˝ C. All measurements were carried out in a sealed chamber at 25 ˝ C under atmosphere controlled with a constant flow rate (0.5 L/min) of synthetic air [44]. 2.3.3. Gas Measurements in Photo-Activation Mode For the electrical characterization in photo-activation mode, the sensors were placed in a dedicated chamber provided with a glass window, through which the light emitted by a Light Emitting Diode (LED) was focused onto the films by means of an optical system (Figure 1b). The LEDs were quasi-monochromatic with an emission spectrum width of 5 nm. Synthetic air and gases were injected into the chamber by the flow-through technique. The electrical conductance of the films was recorded
Sensors 2016, 16, 296
6 of 19
continuously during the experiments by using a simple circuit based on an operational amplifier. The films were maintained under a continuous flow of synthetic dry air during all experiments. At first, CdS and SnS2 sensors were tested in dark condition. In this case, we observed that the conductance was too low to be measured with our experimental setup (
